Kipod FaceTrack Module Uses Machine Learning Technology
Kipod’s face recognition module is based on deep learning neural networks and is able to identify and verify a person and notify the user if any alarm event takes place. Such system greatly enhances security and improves monitoring accuracy.
Nowadays artificial intelligence is used in a number of spheres, including security. Based on machine learning, the industry is able to develop such modules as face recognition, number-plate recognition and audio recognition. One of the most deployed is facial recognition module, which is a great asset for operators and
The face recognition processing flow seems to be quite simple. First, the face is detected, then aligned, then certain features are extracted and finally the face is matched to the reference from the database. At Kipod, we use deep learning network architectures for developing our module, same method as Facebook is using to recognize user faces. Due to local receptive field, common weights and hierarchical organization with spatial subsampling, convolutional neural network ensures a relatively stable resistance to the changes in scale, shifts, turns, changes of angles.
The key features of Kipod FaceTrack module incude:
- Cutting edge neural network architecture with deep learning
- Distributed in-memory database for face matching
- Face feature detector (sex, age group, race, hat, eyeglasses, beard, mustache)
One of the most popular application for face recognition module is security purpose. Due to its high accuracy, it lowers the amount of alarm events for the operator and increases the work efficiency. The system provides accurate screening and algorithms can be set to either notify the user about unrecognized face or, on the opposite, send notification if a face is recognized. The latter is especially beneficial for security services (police, custom service) in order to find a suspect or criminal.
Face recognition module is also successfully used in public places, like stadiums. The system identifies the visitors who bought the tickets (“white list”) and the ones who are in the trespassers database (“black list”). The module enhances security at public places and provides safety for the visitors, being able to immediately detect a person who is listed as trespasser and notify the operator about it.